Don Hedeker Department of Public Health Sciences ... · Singer & Willett (2003) Applied Longitudinal Data Analysis, Oxford University Press Allison (1995) Survival Analysis using

Mixed Models for Discrete- and Grouped-TimeClustered Survival Data

Don HedekerDepartment of Public Health Sciences

Biological Sciences DivisionUniversity of Chicago

hedeker@uchicago.edu

This work was supported by National Institute of Mental Health Contract N44MH32056.

1

Modeling time until an event occurs

• initiation of smoking experimentation in adolescents

• time until school suspension in “problem” kids

• time until start (or end) of service use

• time until quit or relapse (smoking, alcohol, drugs, weight)

• time until death

analysis is called “survival” analysis, but why be so morbid?

⇒ it can be used for any time-to-event data

2

Metric of time

• Continuous time - event timing is known in fine detail

– days until disease development (or recovery)

• Grouped time - event timing is known within intervals of time(also called interval-censored)

– smoking initiation assessed yearly from 7th to 10th grades

• Discrete time - event timing is known, but discrete number oftimepoints and no time intervals

– person failed on the 5th question in a TV game show

Focus on grouped- and discrete-time, but continuous time can bemodelled similarly (using, say, 10 quantiles for event-timeintervals, see Liu & Huang, Statistics in Medicine, 2008)

3

Reading materials - no random effects

• Singer & Willett (2003) Applied Longitudinal DataAnalysis, Oxford University Press

• Allison (1995) Survival Analysis using the SAS System: APractical Guide

• Xie, McHugo, Drake, & Sengupta (2003). Using discrete-timesurvival analysis to examine patterns of remission fromsubstance use disorder among persons with severe mentalillness. Mental Health Services Research, 5, 55-64.

4

Reading materials and examples - with random effects

• Hedeker, Siddiqui, & Hu (2000). Random-effects regressionanalysis of correlated grouped-time survival data. StatisticalMethods in Medical Research, 9:161-179available via www.uic.edu\∼hedeker

• Hedeker & Mermelstein (2011). Multilevel analysis of ordinaloutcomes related to survival data. Handbook of AdvancedMultilevel Analysis, (pp. 115-136), Hoop & Roberts (eds.),Taylor and Francis.

• SuperMix www.ssicentral.com/supermix/downloads.html

– www.ssicentral.com/supermix/examples/Survival.html

– in Supermix (even the free student version), from Help menu, select“Contents,” “Examples from SMIX manual,” “Grouped- anddiscrete-time survival data”

5

Notation is our friend!

• i = 1, . . . , N level-2 units (clusters or subjects)

• j = 1, . . . , ni level-1 units (subjects or multiple failure times)

• assessment time takes on discrete positive valuest = 1, 2, . . . ,m representing time points or intervals

• each ij unit is observed until time tij

– an event occurs (tij = t and δij = 1)

– observation is censored (tij = t and δij = 0)

• censoring: unit is observed at tij but not at tij + 1

• δij is the censor/event indicator

⇒ Outcome is tij (which is either censored or not)

6

Failure, Survival, and Hazard probabilities

cumulative Failure probability, up to and including time t

P (tij) = Pr(tij ≤ t)

cumulative Survival probability beyond time t

1− P (tij)

Hazard = conditional probability that an event occurs at time tgiven that it has not already occurred

p(tij) = Pr(tij = t | tij ≥ t) = (# events at t) ÷ (# at risk at t)

⇒ “ time-interval t” instead of “time t” for time-interval data

7

Kaplan-Meier Survival Function estimates

Initiation of smoking experimentation in adolescents

interval cumulativetime # censor # event hazard prob survival prob survival prob

Females (N=814)

post-int 105 130 130814 = .160 .840 .840

year 1 154 117 117814−235 = .202 .798 (.840)(.798) = .671

year 2 229 79 79814−235−271 = .257 .744 (.671)(.744) = .499

Males (N=742)

post-int 83 156 156742 = .210 .790 .790

year 1 134 89 89742−239 = .177 .823 (.790)(.823) = .650

year 2 217 63 63742−239−223 = .225 .775 (.650)(.775) = .504

8

9

Categorical Regression Models - right-hand side

γt + x′ijβ + z′ijυi

• γt represent baseline hazard

• xij are covariates

– at level-1, level-2, or cross-level interactions

– can include polynomials, dummy variables, interactions, ...

• β are the regression coefficients for the covariates

• zij are the random effect variable(s)

– usually just an intercept for clustered data

– often an intercept and time for longitudinal data

• υi are the random effects ∼ N(0, Συ)

– how cluster i influences the observations within the cluster

– how a subject starts and progresses across time

10

Discrete or Grouped Time?

Discrete time: events occur at discrete points in time

• repeated tasks, e.g., Who wants to be a millionaire?

• logit link: discrete-time proportional odds model

log

P (tij)

1− P (tij)

= γt +[x′ijβ + z′ijυi

]

• with no random effects, same as TIES=DISCRETE option inSAS PROC PHREG in terms of β

• + in formulation means as β ↑ event occurs sooner(i.e., hazard is increased)

11

Grouped time: events occur within continuous time intervals(also called interval-censored time)

• grades of school, e.g., smoking initiation in past year

• complementary log-log link: underlying proportional hazardsmodel in continuous time

log[− log(1− P (tij))

]= γt +

[x′ijβ + z′ijυi

]

• with no random effects, same as TIES=EXACT option inSAS PROC PHREG in terms of β

• + in formulation means as β ↑ event occurs sooner(i.e., hazard is increased)

12

Logit or clog-log link?

• very similar results (so, in practice, it doesn’t matter)

• logit yields odds ratio interpretation for exp β

– logit has proportional odds assumption

• clog-log yields hazards ratio interpretation for exp β

– clog-log has analogous proportional hazards assumption ascontinuous-time Cox model

• clog-log most useful for grouped-time

– where time is really continuous, but measurement onlyoccurs at discrete timepoints and captures eventinformation about a time interval

• logit most common for discrete-time

– no advantage for clog-log over logit for truly discrete-time

13

Initiation of smoking experimentation in adolescents

interval interval interval hazard oddshazard survival odds ratio ratio

time p 1− p p/(1− p) (M/F) (M/F)

Females (N=814)

post-int .160 .840 .190

year 1 .202 .798 .253

year 2 .257 .744 .345

Males (N=742)

post-int .210 .790 .269 1.313 1.416

year 1 .177 .823 .215 .876 .850

year 2 .225 .775 .290 .875 .841

Hazard ≈ odds if p is small (rare event)

14

Two ways to structure the data and analyses

• Ordinal

– ordinal representation of survival time

– analysis using ordinal regression models

– logit or clog-log in terms of P (tij) (cumulative failure)

• Binary

– creation of “person period” indicator(s) for eachobservation to represent survival time

– analysis using binary regression models

– logit or clog-log in terms of p(tij) (hazard)

⇒ Ordinal is easier in terms of dataset structure, but binary iseasier (and more general) in terms of analysis

15

Survival data as categorical outcomes

Ordinal: 2 (post-baseline) timepts with no intermittent censoring

• Outcome = 1 : died at T1 (interval between T0 and T1)

• Outcome = 2 : died at T2 (interval between T1 and T2)

• Outcome = 3 : did not die at T2 (censored at T2)

16

Dichot: 2 (post-baseline) timepts with no intermittent censoring

Create person-time indicators y1 & y2 (0=censor, 1=event)# of records depends on timing of event “person-period dataset”

• y1=1: died at T1 (interval between T0 and T1)

• y1=0 and y2=1: died at T2 (interval between T1 and T2)

• y1=0 and y2=0: did not die at T2 (was censored at T2)

17

Three timepoints with censoring

Ordinal Dichotomousordinal event up to 3 records

outcome dep var indicator per personCensor at T1 1 0 y1=0Event at T1 1 1 y1=1Censor at T2 2 0 y1=0

y2=0Event at T2 2 1 y1=0

y2=1Censor at T3 3 0 y1=0

y2=0y3=0

Event at T3 3 1 y1=0y2=0y3=1

lower values of the ordinal dependent variable signify “worse” outcome

18

Dichotomous or Ordinal representation?

• Results are the same or similar

– clog-log link: identical results for proportional hazardsestimates (i.e., effects that don’t vary with time)

– logit link: similar results

• Ordinal is more efficient in terms of dataset size, especially asnumber of timepoints is large

• Dichotomous more easily allows inclusion of time-dependentcovariates and non-proportional hazards (or odds) models

– each person has a record for each pertinent timept, soinclusion of time-dependent covariate is easy

19

e.g., for a subject with three timepoints of data:

time- time-invariant dependent time

outcome covariate covariate indicatorsy1=0 sex intentions1 0 0y2=0 sex intentions2 1 0

y3=0 or =1 sex intentions3 0 1

• values of intentions change across time

• adding covariate interactions with time indicators allowassessment of proportional hazards (odds) assumption

– without interactions: proportional hazards (odds)

– with interactions: non-proportional hazards (odds)

20

Decisions, decisions ..

data representationlink dichotomous ordinallogitclog-log

• don’t sweat it, results are the same or very similar, which iswhy many prefer dichotomous & logit combination

• for grouped-time data, clog-log would seem to be best choice(in agreement with Cox proportional hazards model forcontinuous time)

• any interest in non-proportional effects or time-dependentcovariates, then dichotomous representation is best

21

School-based Smoking Prevention StudyThe Television School and Family Smoking Prevention andCessation Project (Flay, et al., 1988);

• sample - 2952 7th-graders - 135 classrooms - 28 schools in LosAngeles area

• outcome

– “Have you ever tried a cigarette? (yes/no)”

• timing - students assessed at

– pre-intervention (1/86) (n = 1556 never tried)

– post-intervention (4/86)

– 1 year follow-up (4/87)

– 2 year follow-up (4/88)

22

• design - schools randomized to intervention conditions,interventions delivered in classrooms

– a social-resistance classroom curriculum (CC)

– a media (television) intervention (TV)

– CC combined with TV

– a no-treatment control group

Question of interest:

• Intervention effect on smoking initiation at post-interventionand 2 yearly follow-ups?

23

24

Four timepoints, but first is missing or excludedOrdinal - c:\SuperMixEn Examples\Workshop\Survival\SmkCCLC.ss3

Dichotomous - c:\SuperMixEn Examples\Manual\Survival\SmkBCD2.ss3

Ordinal Dichotomousordinal event (up to 3 records per person)

outcome dep var indicator dep var time indicatorsCensor at baseline 1 0 not in datasetEvent at baseline 1 1 not in datasetCensor at post-int 2 0 y1=0 0 0Event at post-int 2 1 y1=1 0 0Censor at 1 yr 3 0 y1=0 0 0

y2=0 1 0Event at 1 yr 3 1 y1=0 0 0

y2=1 1 0Censor at 2 yr 4 0 y1=0 0 0

y2=0 1 0y3=0 0 1

Event at 2 yr 4 1 y1=0 0 0y2=0 1 0y3=1 0 1

25

Grouped-Time Onset of Cigarette Experimentation in 1556 studentsProportional Hazards Model estimates (se)

PROC PHREG clog-log regressionterm (ties=exact) dichot ordinalintercept β0 -1.652 -1.652

(.091) ( .091)intercept β0 + γ2 -1.613 -.939

(.096) (.083)intercept β0 + γ3 -1.344 -.428

(.106) (.081)

Male β1 .056 .056 .056(.080) (.080) (.080)

CC β2 .041 .041 .041(.080) (.080) (.080)

TV β3 .023 .023 .023(.080) (.080) (.080)

−2 logLfull model 3166.7 3187.4 3187.4with β2 = β3 = 0 3167.0 3187.8 3187.8

26

Grouped-Time Onset of Cigarette Exp. - 1556 students in 28 schoolsMixed-effects Proportional Hazards estimates (se)

term Dichot Ordinalintercept β0 -1.657 -1.657

(.095) ( .095)intercept β0 + γ2 -1.617 -.944

(.101) (.087)intercept β0 + γ3 -1.346 -.432

(.111) (.085)

Male β1 .058 .058(.080) (.080)

CC β2 .045 .045(.084) (.084)

TV β3 .021 .021(.084) (.084)

School variance σ2υ .0031 .0031[r = .002] (.011) (.011)

−2 logLfull model 3187.4 3187.4with β2 = β3 = 0 3187.7 3187.7

27

Ordinal representation - c:\SuperMixEn Examples\Workshop\Survival\SmkCCLC.ss3

28

29

30

31

32

33

Testing of proportional hazards assumption

Relatively easy in dichotomous formulation by includinginteractions with time indicators, e.g., for a subject with threetimepoints:

time timeoutcome covariate indicators interactionsy1=0 sex 0 0 sex × 0 sex × 0y2=0 sex 1 0 sex × 1 sex × 0

y3=0 or y3=1 sex 0 1 sex × 0 sex × 1

Likelihood ratio test: compare deviances (-2 log L) from twomodels, where one is nested within the other. Smaller deviancevalues are better, and the difference can be compared to a χ2

distribution with q df (q = # of additional parameters in largermodel)

34

In present case:

term likelihood-ratio χ2 df p <intervention (CC & TV) 4.1 4 nssex 8.0 2 .02

From model with sex by time interaction terms:

term estimate std error z-statistic p <Male at Post-Int .306 .119 2.57 .011Male by Year 1 -.452 .184 -2.46 .015Male by Year 2 -.458 .207 -2.21 .028

Male at Year 1 -.146 .141 -1.03 nsMale at Year 2 -.152 .170 -.89 ns

35

Grouped-Time Onset of Cig. Exp. - 1556 students in 28 schoolsMixed-effects Partial Proportional Hazards estimates (se)

term estimate std error p <Intercept -1.784 .108 .001

Year 1 .260 .128 .042

Year 2 .536 .143 .001

Sex (f=0; m=1) .306 .119 .011

CC (no=0; yes=1) .047 .084 .576

TV (no=0; yes=1) .021 .083 .805

Sex × Year 1 -.452 .184 .015

Sex × Year 2 -.458 .207 .028

School variance .0029 .011 .788

36

Binary representationc:\SuperMixEn Examples\Manual\Survival\SmkBCD2.ss3

37

38

39

40

41

42

43

44

45

46

47

48

49

Gender effect - estimated hazard ratios

• post-intervention: exp(.3059) = 1.36 ⇒ Males hazard ofsmoking is significantly increased (an increase of about 36%)

• year 1: exp(−.1458) = .86 ⇒ Males hazard of smoking isreduced (about 14%), but not significant

• year 2: exp(−.1517) = .86 ⇒ Males hazard of smoking isreduced (about 14%), but not significant

note: these estimates are conditional estimates accounting forschool, CC, and TV effects

50

Kaplan-Meier Survival Function estimates

Initiation of smoking experimentation in adolescents

interval cumulativetime # censor # event hazard prob survival prob survival prob

Females (N=814)

post-int 105 130 130814 = .160 .840 .840

year 1 154 117 117814−235 = .202 .798 (.840)(.798) = .671

year 2 229 79 79814−235−271 = .257 .744 (.671)(.744) = .499

Males (N=742)

post-int 83 156 156742 = .210 .790 .790

year 1 134 89 89742−239 = .177 .823 (.790)(.823) = .650

year 2 217 63 63742−239−223 = .225 .775 (.650)(.775) = .504

51

Model fit of response proportionsPartial Proportional Hazards (random schools) model - Dichotomous

Sex clog-log Ψ(z) = 1− exp(− exp(z)) est.

Hazard probability at Post-Int

F Ψ((−1.785 + .47× .047 + .48× .021)/√d̂) .159

M Ψ((−1.785 + .306 + .47× .047 + .48× .021)/√d̂) .210

Hazard probability at Year 1

F Ψ((−1.785 + .261 + .47× .047 + .48× .021)/√d̂) .202

M Ψ((−1.785 + .306 + .261− .452 + .47× .047 + .48× .021)/√d̂) .176

Hazard probability at Year 2

F Ψ((−1.785 + .536 + .47× .047 + .48× .021)/√d̂) .257

M Ψ((−1.785 + .306 + .536− .458 + .47× .047 + .48× .021)/√d̂) .225

d = design effect = (σ2υ + σ2)/σ2 d̂ = (.0029 + π2/6)/(π2/6)

.47 = CC mean, .48 = TV mean

52

Model Fit

53

Youth within therapists example

Schoenwald, S.K. (2008). Toward evidence-based transport ofevidence-based treatments: MST as an example. Journal ofChild and Adolescent Substance Abuse, 17(3), 69-91.

“has child been suspended in the current school year”

visit 1 visit 2 visit 3 visit 4no 1089 1122 1074 1046yes 783 611 445 335

visit 1 = baseline, visit 2 = post-int, visit 3 = 6-months, visit 4 = 12-months

outcome of interest: time until first school suspensioncovariates: child gender, family financial assistance

54

• 1914 youth nested within 443 therapistsCumulative Cumulative

n Frequency Percent Frequency Percent

1 107 24.15 107 24.15

2 85 19.19 192 43.34

3 51 11.51 243 54.85

4 43 9.71 286 64.56

5 35 7.90 321 72.46

6 27 6.09 348 78.56

7 26 5.87 374 84.42

8 14 3.16 388 87.58

9 10 2.26 398 89.84

10 6 1.35 404 91.20

11 10 2.26 414 93.45

12 6 1.35 420 94.81

13 6 1.35 426 96.16

14 7 1.58 433 97.74

15 4 0.90 437 98.65

16 2 0.45 439 99.10

17 1 0.23 440 99.32

19 2 0.45 442 99.77

26 1 0.23 443 100.00

55

c:\SuperMixEn Examples\Primer\Survival\Suspend.ss3

56

57

58

59

60

61

62

Kaplan-Meier Survival Function estimates

Time to first school suspension

# # hazard interval cumulativetime censor event prob surv prob survival prob

Males with financial assistance (N=473)

baseline 14 223 223473 = .471 .529 .529

post-int 26 69 69(473−237) = .292 .708 (.529)(.708) = .374

6-months 13 30 30(473−237−95) = .213 .787 (.374)(.787) = .294

12-months 83 15 15(473−237−95−43) = .153 .847 (.294)(.153) = .249

⇒ Similar calculations for other groups (males withoutassistance, females with assistance, females without assistance)

63

64

Model fit - Males with financial assistanceProportional Hazards (random therapists) model - Ordinal

clog-log Ψ(z) = 1− exp(− exp(z)) estimate (1 - estimate)∗

Probability of Category 1 response: Failure at Baseline

Ψ((−.656 + .200)/√d̂) = .470 .530

Prob of Category 1 or 2 response: Cumulative Failure at Post-Int

Ψ((−.224 + .200)/√d̂) = .624 .376

Prob of Category 1, 2, or 3 response: Cum Failure at 6-months

Ψ((−.032 + .200)/√d̂) = .694 .306

Prob of Category 1, 2, 3, or 4 response: Cum Failure at 12-months

Ψ((.121 + .200)/√d̂) = .748 .252

d = design effect = (σ2υ + σ2)/σ2 d̂ = (.0834 + π2/6)/(π2/6)

∗ (cumulative) survival = 1 - cumulative failure estimates

65

Model Fit

66

Model without Sex by Financial Assistance

comparing models with and without interaction, vialikelihood-ratio test, χ2

1 = 4741.49696− 4741.46612 = .03

variable estimate std error z-value p-valueSexF -0.3293 0.0654 -5.0362 0.0000FinAsst 0.1933 0.0621 3.1109 0.0019

exp(−.3293) = .719 ⇒ Females hazard of school suspension issignificantly reduced (a reduction of about 28% relative to males)

exp(.1933) = 1.213 ⇒ Financial assistance kids havesignificantly increased hazard (an increase of about 21%)

note: these estimates are conditional estimates, accounting forthe therapist effects

67

Conditional vs Marginal effects

• In a mixed model, the regression coefficients and the randomtherapist effects are jointly estimated

• regressor effects are obtained controlling for, or adjusted for,or conditional on the therapist effects

– comparing the populations of boys versus girls, controllingfor therapists (i.e., how different are the populations ofboys and girls who have the same therapist)

• marginal effects or unconditional effects are sometimes of(greater) interest (i.e., population-averaged effects)

– comparing the populations of boys versus girls

• in linear mixed models, conditional = marginal effects, butthis is not true, in general, in non-linear mixed models (i.e.,mixed models for non-normal outcomes)

68

Expressing conditional as marginal effects

In a random intercept model, βM = βC /√d

• βM and βC are the marginal and conditional effects

• d is the design effect = (σ2υ + σ2)/σ2

in current example, d = (.0834 + π2/6)/(π2/6) = 1.0507

−.3293/√

1.0507 = −.3213 marginal sex effect

.1933/√

1.0507 = .1886 marginal financial assistance effect

exp(−.3213) = .725 ⇒ Females hazard of school suspension issignificantly reduced (a reduction of about 27% relative to males)

exp(.1886) = 1.208 ⇒ Financial assistance kids havesignificantly increased hazard (an increase of about 21%)

69

Degree of clustering attributable to therapists

Calculation of the intracluster correlation

residual variance = pi*pi / 6 (assumed)

cluster variance = 0.0834

intracluster correlation = 0.0834 / ( 0.0834 + (pi*pi/6)) = 0.048

⇒ fair degree of clustering within therapists

• suggests that some therapists have positive effect on time toschool suspension, others have negative effect

70

Empirical Bayes estimates of random effects

log[− log(1− P (tij))

]= γt+x

′ijβ+υi where υi ∼ N(0, σ2

υ)

• Random effects υi are also estimated

• can be of interest to indicate how particular clusters (i.e.,therapists) are doing

• can be used to rank or compare clusters, or indicate unusualclusters

• SuperMix provides these under “Analysis,” “View level-2Bayes results” (also saved as a file with .ba2 extension)

• graph them under “File,” “Model-based Graphs,”“Confidence Intervals”

71

ID, random effect number, random effect estimate (standardized θi = υi/συ),(posterior) variance, random effect label

72

θ̂i ± 1.96√

therapist’s posterior variance

73

SAS for reading in Empirical Bayes estimates

DATA one;

INFILE ’c:\SuperMixEn Examples\Primer\Survival\Suspend1.ba2’;INPUT id r1 TherInt TherPrec intercpt $;

PROC SORT; BY TherInt;

PROC PRINT; VAR id TherInt TherPrec;

RUN;

Obs id TherInt TherPrec

1 265 -0.35481 0.047210

2 354 -0.34406 0.049831

3 123 -0.33236 0.062671

4 122 -0.32261 0.059428

. . . .

. . . .

440 175 0.32769 0.059300

441 400 0.36221 0.061400

442 61 0.36267 0.055196

443 173 0.36696 0.052603

74

And the winner is ...

Therapst YouthID Suspend Event SexF FinnAsst SexFin265 422 1 0 0 0 0265 510 4 0 1 0 0265 572 3 0 0 0 0265 594 4 0 0 0 0265 640 1 1 0 1 0265 747 1 1 0 1 0265 1101 3 0 0 0 0265 1340 2 1 0 1 0265 1505 3 1 0 1 0265 1667 4 0 0 1 0265 1863 3 0 0 0 0265 1926 4 0 0 0 0265 2011 4 0 0 1 0265 2016 3 1 0 1 0

mostly censored observations with higher times to first suspension

75

And the loser is ....

Therapst YouthID Suspend Event SexF FinnAsst SexFin173 200 1 1 0 0 0173 279 1 1 0 0 0173 382 2 0 1 1 1173 477 2 1 1 0 0173 523 1 1 0 0 0173 760 1 1 0 1 0173 923 1 1 0 0 0173 1242 1 1 0 1 0173 1610 1 1 0 0 0173 1646 1 1 0 0 0173 1725 2 0 1 0 0173 1795 1 1 1 1 1173 1991 4 0 1 0 0173 2013 1 1 0 0 0173 2250 1 1 0 0 0

mostly event observations with lower times to first suspension

76

Second thoughts

• Assessing effects of therapists including baseline seemsproblematic

• Being suspended at baseline seems unrelated to therapisteffectiveness

• some therapists might be getting more (or less) kids withbaseline suspension

• seems reasonable to exclude baseline, and focus on time tofirst suspension after baseline

77

Excluding baseline visit

78

79

80

Degree of clustering attributable to therapists

Calculation of the intracluster correlation

residual variance = pi*pi / 6 (assumed)

cluster variance = 0.0010

intracluster correlation = 0.0010 / ( 0.0010 + (pi*pi/6)) = 0.001

⇒ very small degree of clustering within therapists

81

SAS for reading in NEW Empirical Bayes estimates

DATA two;

INFILE ’c:\SuperMixEn Examples\Primer\Survival\Suspend2.ba2’;INPUT id r1 TherInt TherPrec intercpt $;

PROC SORT; BY TherInt;

PROC PRINT; VAR id TherInt TherPrec;

RUN;

Obs id TherInt TherPrec

1 122 -0.17915 0.056097

2 211 -0.17415 0.056336

3 354 -0.14976 0.051612

4 103 -0.14740 0.051710

. . . .

. . . .

388 481 0.18269 0.061248

389 482 0.21592 0.063285

390 238 0.21776 0.059572

391 610 0.26182 0.058515

82

And the NEW winner is ...

Therapst YouthID Suspend Event SexF FinnAsst SexFin122 243 3 0 1 0 0122 391 4 0 1 0 0122 531 4 0 0 0 0122 576 4 0 0 0 0122 577 3 0 0 0 0122 704 3 0 1 0 0122 705 4 0 1 0 0

And the NEW loser is ...

Therapst YouthID Suspend Event SexF FinnAsst SexFin610 1291 4 1 1 0 0610 1371 2 1 0 0 0610 1728 4 0 1 0 0610 1740 2 1 0 1 0610 2082 2 1 0 1 0610 2188 2 1 0 0 0

83

Model without Sex by Financial Assistance

comparing models with and without interaction, vialikelihood-ratio test, χ2

1 = 2194.86989− 2194.40487 = .565

variable estimate std error z-value p-valueSexF -0.3223 0.1027 -3.1391 0.0017FinAsst 0.2026 0.0999 2.0266 0.0427

exp(−.3223) = .725 ⇒ Females hazard of school suspension issignificantly reduced (a reduction of about 27% relative to males)

exp(.2026) = 1.225 ⇒ Financial assistance kids havesignificantly increased hazard (an increase of about 23%)

note: these estimates are conditional estimates, accounting forthe (near-zero) therapist effects

84

Tests of proportional hazards assumption

In ordinal, fit models with and without “Explanatory VariableInteractions” on Advanced card

term likelihood-ratio χ2 df p <financial assistance 3.45 2 nssex 2.03 2 ns

85

Summary

• Time-to-event analysis for clustered (or repeated) discrete-and grouped-time data

– dichotomous or ordinal mixed regression models

• Extenstions to competing risk survival models (Gibbons et al,2003, Biostatistics)

– person-time indicators (0=no event or censoring, 1=eventA, 2=event B)

– nominal (mixed) regression model

• Can also be used for continuous-time analysis (groupingtime-to-event outcomes in, say, 10 quantiles of time periods)

– lack of software for continuous-time (mixed) analysis

– Liu & Huang, (Stat Med, 2008) provide simulation resultssupporting this approach

86